All means will be welcome to help data flow across smart cities. As a matter of fact, smart cities widely rely in sensing the environment, an action that is prone to generating huge amounts of data. A major challenge is how to collect such data in an efficient way without the need to deploy, whenever necessary, extra (costly) cellular infrastructure. In this paper, we examine the possibility of creating a delay-tolerant vehicular network in the city of Rio de Janeiro, Brazil, using the public transportation system as a data carrier. We evaluate the capacity of such a network by analyzing a large mobility dataset reporting GPS positions of 12,456 buses and 5,833 taxis during a 24-hour period. Our results confirm the viability of the approach and reveal that hundreds of Terabytes can circulate across the city on a daily basis while achieving significant city coverage.